« Back to Volume 14 (2005), Issue 3
Pojdi na slovensko stran članka / Go to the article page in Slovene
Statistical significance and its meaning
Christina Bachmann, Riccardo Luccio & Emilia Salvadori
Full text (pdf) | Views: 19 | Written in Slovene. | Published: December 15, 2005
Abstract: The most frequently method of data analysis in psychology is null hypothesis significance testing (NHST); this work identifies the misunderstandings and consequent biases due to the unwitting use of this hybrid approach, derived by the fusion of the p-value approach (PVA) of Fisher (1925) and the fixed alpha approach (FAA) of Neyman and Pearson (1933). The aim of this work is to contribute to the critical debate about the use of this test, how researchers interpret test results, and the logic of the NHST approach. In the same vein of recent recommendations of the APA Task Force, we examine different methods of data analysis, that are used with NHST including confidence intervals, effect size measures, power analysis, and resampling.
Keywords: null hypothesis testing, statistical significance, power analysis, effect size (statistical)
